It is said that structural analysis of human genome will be completed in or before 2003 as the human genome project is progressing well. Now, the age of isolating genes one by one and analyzing their structures separately seems to be over, and we have come into the age of “structural analysis” of genome.
With the nucleotide sequence of genome alone, however, information on functions is insufficient. Thus, a novel analysis system for functional analysis is needed. Further, although one of the major goals of human genome analysis is to elucidate causative genes in human diseases, such diseases cannot be explained with the structures of causative genes alone.
Accordingly, production of model individuals is an indispensable assignment in order to analyze processes of disease development and to develop new treatment methods after the identification of causative genes.
On the other hand, if genome is divided into gene regions and non-gene regions in terms of structure, it is considered that these two parts have separate functions and that it is necessary  to analyze the functions of both parts (FIG. 1 and Supplemental FIG. 1). From the viewpoint of entire genome, each gene is performing only a part of the entire function. Genome is not a mere collection of genes and may have unknown functions. In fact, a new concept “position effect mutation” has been established. From this, it is presumed that genome has regions of unknown functions.
Gene regions are composed of regulatory regions and coding regions. At present, the target of genome functional analysis is coding regions. When mouse is compared with human, the kinds of genes they have are almost equal. Therefore, functional analysis of the regulatory region is important. There is difference in species between mouse genes and human genes. It is believed that this difference is not due to difference in protein but due to difference in the regulation of gene expression.
The function of a transcription factor or the like involved in the regulation of gene expression can be elucidated from the sequence of the coding region of the relevant gene. The analysis of the functions of those elements to which the transcription factor binds is extremely difficult at present because a number of those elements exist in the regulatory region of one gene. However, as a technique of functional analysis, a method using bacterial artificial chromosomes may be considered.
It is considered that functional analysis of coding regions may be performed at the mRNA level, protein level, cell level, tissue/organ level and individual (i.e. whole animal) level. It is believed that such analysis at the mRNA level can be performed using DNA chips. On the other hand, the use of embryonic stem (ES) cells seems to be the best way for performing functional analysis at other levels, because various cell and tissue derivative systems have been developed directly from ES cells in vitro and a number of such systems are expected to be developed in the future. Furthermore, the use of ES cells is advantageous in that individual level analysis systems can be established.
From the foregoing, it is understood that gene knockout at ES cell level and production of knockout mice in which the relevant genes are knocked out are extremely important in functional analysis of genome.
To date, homologous recombination using ES cells has played a major role in the production of knockout mice. However, considering this method not as a strategy of producing knockout mice separately but from a strategic viewpoint of producing knockout mice comprehensively, this method has serious problems.
First, this method requires too much time. In the production of knockout mice, it is the rate-determining step to isolate knockout ES clones generated through homologous recombination using ES cells. Even a skillful researcher needs at least three months for isolating a knockout ES clone. Thus, only four genes can be knocked out in one year. Accordingly, in the case of introducing each one mutation into 105 genes, 2,500 researchers are required for one year. It is estimated that approximately 1,000 lines of knockout mice are produced in one year in the world. This means that it would take 100 years to produce 105 knockout ES clones. This is so unrealistic compared to the advance in the structural analysis of human genome that is to be completed in 2003.
Secondly, this method requires too much cost. At least 2 to 4 million yen is necessary to produce one line of knockout mouse excluding personnel expenses and depreciation expenses. Thus, production of 105 simple knockout mice requires 200 to 400 billion yen.
As described above, the conventional homologous recombination using ES cells has problems, and genome is vast. However, the number of genes in genome is limited. Thus, it is necessary to isolate from genome those genes having important functions. In many cases, the function of a gene is elucidated only after production of a knockout mouse in which the relevant gene is disrupted. Therefore, knockout mice are directly connected with future development of epoch-making drugs and have extremely high value added. Under circumstances, it has become the world's “strategy” to produce mutant mice at random and in large scale. At present, the three methods described below are considered most reasonable in the production of random mutation mice.
The first one is a method using ethyinitrosourea (ENU), a mutagen. A project of large-scale mutant production using ENU has been started in Europe. In Germany, Dr. Balling of the Institute of Mammalian Genetics and others started this project in 1997 as a part of the human genome project. In England, supported by SmithKline Beecham, Dr. Brown and others started this project at MRC Mouse Genome Center in Harwell aiming at establishment of mutant mice having mutations mainly in brain/nervous system. To date, these two groups have established approximately 200 lines of mutant mice exhibiting dominant inheritance. The project is proceeding more efficiently than expected. In the United States, it has been decided that structural analysis of mouse genome and production of mutant by the ENU method start with a huge budget (6 billion yen/year) at Case Western Reserve University, Oak Ridge National Laboratory, etc.
When ENU is administered to adult male mice, ENU acts on spermatogoniums before meiosis and causes about 50 to 100 point mutations per spermaogonium at random. Mutations occur at a frequency of 1/1,000/gamete per locus. Therefore, by crossing one treated male mouse with one normal female mouse, many kinds of mutant mice can be produced in F1 generation. In the method using ENU, if 1,000 mice are screened for a specific locus, one mouse has a mutation caused in that locus in terms of probability. Thus, this method is considered highly efficient.
The second method is a method using chlorambucil that is also a mutagen. This method causes mutations in spermatogoniums at the same frequency as in the method using ENU. However, these mutations are deletion mutations, and sometimes as many as one megabases may be deleted.
The third method is a method using gene trapping. Gene trapping is a technique that was developed for the purpose of searching for unknown genes by introducing trap vectors containing a marker gene into ES cells and then monitoring the expression of the marker gene. Trap vectors are integrated into ES cells at random and, as a result of their integration, endogenous genes (genes present in cells and tissues inherently) are disrupted in most cases. Therefore, preparing chimeric mice from such ES cells can produce various knockout mice.
However, each of the methods using a mutagen and the method using gene trapping has an advantage(s) and a drawback(s) (Table 1).
TABLE 1ENUChlorambucilMethodMethodGene TrappingNature of MutationPointDeletion mutationAny desiredmutationmutationProduction of MutantEasyEasyDifficultMouseIdentification ofDifficultMediumEasyMutant MouseOther FeaturesCan use ES trapclones
According to the ENU method, production of mutant mice is easy, but establishment of individual mutant lines is not easy because segregation by crossing should be conducted. Further, in order to identify mutated genes, the relevant locus should be identified first by linkage analysis using polymorphic DNA markers, and then the gene should be isolated by positional cloning. Thus, the ENU method requires complicated operations.
According to the chlorambucil method, production of mutant mice is easy, but deleted sites should be identified. For that purpose, analysis must be made using a number of polymorphic DNA markers. Besides, generally, methods using a mutagen such as chlorambucil need large breeding rooms. Thus, such methods require much expenses and labor.
Although the gene trap method requires labor and technology for producing mutant mice, identification of mutated genes is easy and experiments can be conducted according to the size of breeding rooms. Gene trap ES clones per se are precious resource for functional analysis of genome. The gene trap method is also remarkably different from other methods in this point.
Some laboratories in the world have already started production of mutants by gene trapping. In the United States, a private firm Lexicon Genetics Incorporated is undertaking random disruption by gene trapping using retrovirus vectors. However, ordinary researchers can hardly use this service because of the following reasons. Briefly, it is not sure whether an endogenous gene is disrupted or not even if the gene is trapped; it is not clear whether germline chimeric mice can be produced; an additional charge is required for the production of chimeric mice; and considerable charges are required for using the service. In Germany, gene trapping is performed toward a goal of 12,000 clones as a part of the ENU project. Anyway, these are proceeding focusing on the analysis of trapped genes rather than the establishment of mouse lines.